We were collecting data using a data set having a Monthly period type. the team has decided that we have to change the period type to Weekly. If we change the period type, shall be able to still download the data that was previously entered?
Thank you for your question! It seems like a risky approach if you want to keep the previously entered data, so I’d like to make a suggestion.
Let’s say for example you already have Dataset Input X with period type Monthly and after entering data for three months (or more or less) the team decided to change it to a Weekly period, so my suggestion is to change the sharing settings for Dataset Input A so that it won’t be visible to anyone (and maybe rename it as Dataset Input A (Monthly))
Additionally, instead of changing the current dataset period type, I suggest to create a new dataset with the same name ‘Dataset Input A’ and the new period Weekly to replace the older one and give this one the sharing access settings while revoking access to the previous one.
This way you can keep the data that was entered on a monthly basis, and have a record of the data that will be entered on a weekly basis.
I hope this works for you. Please feel free to post back if you have any concerns or comments.
If you need (or wish) to keep the Monthly dataset for reference, it would also be possible to give the new dataset a different name to distinguish it from the original. (Dataset Input A weekly to follow Gassim’s example.)
I don’t know what kind of data you are collecting, or what kind of analysis and statistics are done before using the information. The one thing with weeks that is problematic is that they don’t correlate with months and years, so aggregation for longer periods can never be accurate. It may be that you actually require data on a daily level of detail, reported once a week. Days as a reporting period of course can be aggregated to Weeks, Months and Years. Days will naturally increase the total size of the analytics database compared to weeks or months, so it can have an impact on performance when running the analytics calculation jobs.
Also please note that indicators with different periods may calculate numbers in ways you may not expect, so check that the output looks sensible and similar to what you used to have.
A final note of caution: Changing the configuration of the data or organisation should always be done in a development or test environment first, and never directly in the production instance of DHIS (or any other system for that matter).
Hi @Gassim and @Paul
I am very happy with your explanation. I may get get back to you if the need arises.